We present a new analytical method for characterizing the directional tuning of neural data. The method is based on computing parameters associated with the geometric properties of solids, and provides an estimate of preferred direction in the context of non-uniform sampling of directions. Unlike optimization methods based on fitting tuning functions, the plate method is computationally fast, and does not require the assumption of an underlying tuning function (e.g. cosine or von Mises functions). In addition to estimating the preferred direction of a dataset, the plate method provides other parameters to fully characterize the directional properties of neural data. The method is presented in the context of a two-dimensional coordinate system but may in principle be extended to higher dimensional spaces as well.